A hydrogen reformer furnace converts natural gas into hydrogen through a series of catalytic reactions. One of the most prevalent routes for the conversion of methane (CH4) to petrochemicals is either through the manufacture of hydrogen, or a mixture of hydrogen and carbon monoxide. This hydrogen/carbon monoxide material is referred to as “Synthesis Gas” or “Syngas.” Indeed, steam methane reforming (SMR) of natural gas or syngas is the most common method of producing commercial bulk hydrogen as well as hydrogen that is used in the industrial synthesis of ammonia. At temperatures around 1000° C., and in the presence of metal-based catalysts, steam reacts with methane to yield carbon monoxide and hydrogen. These two reactions are reversible in nature:CH4+H2OCO+3H2 
The reaction is endothermic and requires the input of large amounts of heat in order to be sustained. Heating gas accounts for 80% of the total process gas requirement.
A common type of hydrogen reformer furnaces is known as a “top down,” or “down fired” furnace. FIG. 1 is a perspective cut-view of a conventional hydrogen reformer furnace 800. Natural gas burners (not shown) are spaced at the top of the furnace 800 in between rows (also known as lanes) of catalyst pipes 70, and produce heat through combustion. The burners fire downward, parallel to the hydrocarbon-steam mixture flow, direction through the catalyst tubes 70, which are centrifugally cast chrome-nickel tubes that are typically 10-40 feet long, and mounted vertically in the furnace 800. The tubes 70 contain an activated nickel catalyst on alumina carriers in the form of pellets or balls, for example. The process gas and steam are fed downward over the catalyst and removed from the bottom of the tubes 70.
The primary reformer operates at temperatures in the 700-800° C. range. The hot gas is then passed into a convective heat transfer zone, and steam is generated and may be fed back into the primary reformer. This unit is used to produce synthetic fuel that may be turned into a variety of liquid fuels for powering internal combustion engines. It is also often used to produce hydrogen for other processes in the plant burner whereby flame and hot gas radiation provide heat input to the tubes to support the highly endothermic reaction. The air exits out one side of the bottom of the furnace 800. Based on the location of the burners and the furnace exit, the air flow and heat distribution are non-uniform. In this situation, it is common for the hot gasses with flow directly to the exit, creating a cold area in the back of the unit and a hot spot just before the exit which can reach temperatures high enough to damage the catalyst tubes. In order to correct for this, flue gas tunnels 80 are provided, which act as collection ducts for the combustion gases, promoting even heat distribution in order to improve efficiency and prolong the life of the tubes 70.
While SMR is a well-established process, and efforts have been made over time to optimize many facets of the technology in order to increase efficiency, most of the focus has been centered on improving aspects of these reformers with respect to the catalyst, metal alloys, burners, feed stocks, etc. However, one piece of SMR technology has been completely ignored where advancements are concerned. That is, the refractory designs used in the construction of these units have remained stagnant for decades. In particular, the flue gas tunnels which transport combustion gasses through the fired heaters have not been changed despite the clear need for improvement based on performance reliability issues.
These tunnels 80 average about 8 feet in height, 3 feet in width and run the full length of the furnace unit 100 (e.g., 40 ft-50 ft). Because of the size of these tunnels 80 and the volume of refractory materials used in their construction, they have traditionally been fabricated using basic brick shapes (e.g., standard rectangular shapes, shown in FIG. 2), in a similar manner to constructing any structural brick wall. The walls 81 of the tunnels 80 are then topped with a series of rectangular blocks 82 that form a lid (see, e.g., FIGS. 1-3). Historically, conventional tunnel walls 81 have been very prone to mechanical failure under heat and over time. The main modes of failure for these flue tunnels are related to refractory design, installation technique, mechanical abuse in service, and initial material selection.
Even though they are problematic, these tunnels 80 are essential in order for the furnace unit 800 heat evenly and achieve the required efficiency. For example, if a down fired reformer did not include such tunnels 80 in its unit construction, all of the combustion gasses would rush into the flue at the exit of the reformer. This would create uneven temperatures throughout the unit with cold areas away from the flue and hot spots near the exit of the unit, as discussed above. As a result, the reformer would not only experience reduced efficiency, but would also overheat the catalyst tubes near the exit, causing premature failure to occur.
The design and construction of conventional flue gas tunnels 80 in SMRs involves the use of flat bricks with typical dimensions of 3 in×9 in×6.5 in. The walls 81 are constructed so that half blocks are left out in regular patterns to allow for gas passage through the wall 81 into the tunnel 80 (not shown). Typically, the bricks are mortared in place during construction in order to hold the walls 81 together. A common alternative to the standard flat brick is a standard tongue and groove brick 83, 84 (see, e.g., FIGS. 4 and 5). While many sizes and configurations of these types of bricks exist, such conventional bricks typically use a simple tongue and groove feature to mechanically engage each other when vertically stacked in the common manner. As shown in FIGS. 4 and 5, conventional bricks 83, 84 include simple tongue 832, 842 and groove-style mating features 833, 843 that fit together when vertically stacked.
In the past, in conventional tunnel structures, large expansion gaps have been provided, located every 6-10 ft along the tunnel walls in order to account for thermal expansion in the system. The expansion gap is a critical aspect of design and construction, because the anticipated thermal growth must be accommodated. In this case, however, due to the presence of these large expansion gaps, every tunnel is actually made up of several large free-standing walls. In order to help support these free-standing sections of tunnel wall, intermediate support walls or pilasters are therefore also provided (not shown). These intermediate support walls connect the outer walls of tunnels between catalyst tubes in order to prevent the walls from leaning or collapsing. Pilasters, also known as buttresses, serve the same purpose, and are structured as columns of bricks located outside of the tunnel walls (not shown).
Another feature of the tunnel wall construction is the end wall (not shown). Also known as cross-over walls or target walls, these brick wall segments connect tunnels at the exit of the unit, preventing gas by-pass through the surrounding lining. In addition to providing additional lateral support, the end walls also ensure that all combustion gasses properly exit through the flue gas tunnels 80.
Once the tunnel walls are constructed, the tunnel covers (lids) are placed on top. These covers, often called coffin covers, are typically made from large slabs of refractory material. However simple the design may be, they serve an important purpose, because failed covers decrease the unit efficiency, cause tunnel wall failure as they fall, and contribute to shorter tube life. There are four main styles of coffin covers. The main style is a rectangular or square solid design (see, e.g., lid 82 in FIG. 3). This represents the traditional approach, and is simply a solid slab of refractory material that spans the horizontal distance (gap) between walls 81. These solid covers 82 can also have a notched surface or otherwise be formed with a mating feature on the bottom or sides that can mechanically engage with the tunnel walls and provide additional support (not shown). Another style is the hollow or extruded lid 821 (see, e.g., FIGS. 6 and 7). These types of covers 821 have the same outer dimensions as the rectangular solid lid 82, but include a pair of hollowed-out sections (cavities) 822 in the middle to reduce the weight of the lid and the resulting stresses.
Another common cover design is the off-set cover 831, as shown in FIG. 8. This solid lid features a slanted geometry that facilitates engagement between adjacent covers, which offers extra support during upsets and which can help support cracked lids in the event of a cover failure. FIG. 9 shows a tongue and groove cover 851, which is a another version of the off-set cover 831, but whose mechanical mating features (i.e., tongue 851a, and groove 851b) provide even more engagement with adjacent lids 851.
One of the current types of failures seen in the field is the collapsing of a section of lids, or all of the lids, over the entire length of the tunnel. Once installed, the lids act as a beam, and a crack in the middle of the lid is often the result of the ratio between the span and the material thickness. The cross-section (thickness) of the replacement lids is then increased, but after another campaign, the failure is typically even worse than before. This is because the lid failure is not a result of static load. Hand calculations coupled with computer simulation have shown that the static load alone imparts very little stress on the lids, and will not result in a failure. Computer run finite element analysis (FEA) of a 9 in W×9 in T×42 in L solid rectangular lid (see, e.g., FIG. 3) installed on a tunnel at a constant service temperature of 1900° F. demonstrated that the lid has no external forces acting upon it other than its own weight. The result is a maximum stress of a very modest 10 psi.
With many materials, the modulus of rupture (MOR) decreases significantly at higher temperatures, and it is possible to select a low grade refractory lid material whose MOR decreases at operational excursion temperatures to a point that even the mild stresses associated with the static load can result in failure. However, most engineered refractory material suppliers characterize the hot modulus of rupture (HMOR), and supply a material option for lids that have a high enough HMOR so that even with the decrease in strength, the static loads still have a very significant factor of safety associated therewith. Based on the comparison of the FEA results to the published HMOR, it has been concluded that most lid failure is not a result of static load alone, and is therefore a result of stresses associated with the thermal state.
Thermal stresses in such a situation manifest several ways. One way the components can fail is if the thermal expansion is not properly managed, resulting in excessive compressional force. Since the lids are placed on top of the wall sections and the only constriction is either friction or mortar, the thermal expansion will not be constricted to the point of failure. The HMOR of commonly used refractory mortars is roughly 500 psi, well below that of the refractory material selected for the tunnel lid, so if the thermal stresses reach that level, the mortar will break and the lid will be free to expand as necessary.
The component can also fail as a result of thermal stress that occurs as a result of any temperature differential incurred during operation, and is not limited to instances of large upsets. Thermal stress failure results when the thermal expansion from one area of a component is different from another area resulting in a stress greater than the yield strength of the material. If the temperature in the convection section of the furnace is different than the temperature inside the tunnels, even for a short period of time, the potential for thermal stress is present. FEA of a 9 in W×9 in T×42 in L solid rectangular lid (see, e.g., FIG. 3) installed on a tunnel with a temperature on the top surface of the lid at 1910° F. and a temperature on the bottom surface of the lid at 1900° F. has shown that the lid has no external forces acting upon it, other than its own weight. A differential temperature of 10 degrees across the lid results in a max stress of 1500 psi, which is above the HMOR of lower end refractory materials. In a situation where a very large number of the lids of a tunnel all failed during the same campaign without any of the walls collapsing, it is most likely that the mode of failure was thermal stress.
Another important factor in the performance of the tunnel lids is the material's creep resistance. Creep occurs when a material slowly but permanently deforms under long term exposure to high levels of stress that are below the material yield strength. The result on the tunnel walls is a transmission of the lids mass in the vertical direction, which compliments the strength and structure of the wall. Creep of a lid will result in a “sagging” of the center span and will change the interaction force between the lid and the tunnel walls, and eventually lead to a failure. Creep can be characterized with ASTM standard testing, which is representative of the use of a tunnel lid in service and is an important component to material selection. ASTM tests on Super Duty Brick have published results of a 7.86% deflection at 2,600° F. The result on the tunnel walls is a transmission of the lids mass at an angle that is a few degrees off of the vertical axis and which encourages the walls to separate further apart from one another at the top than at the bottom.
A full tunnel collapse can actually be the result of several different modes of failure. Conventional tunnel construction uses hundreds of thousands of pounds of refractory brick and lids, all of which accounts for mass that ultimately rests on a final base layer of insulating fire brick (IFB; not shown in FIGS. 1A and 1B). Conventional tunnel cross-sections with bricks that are 6 in wide, tunnel walls that are 96 in tall, and a solid lid that is 9 in thick results in a load on the supporting IFB layer of 11.6 psi. Published data using ASTM testing shows that at the temperatures present in the reformer furnaces, the base IFB layer will deform a full 1% under those loads in 100 hours. The deformation of the base IFB layer translates in one of two ways: either the deformation will prematurely compress the fiber allowances for thermal expansion, or the deformation will reduce the overall insulating value of the base IFB. Both instances are known to result in failure.
The effects of temperature and tunnel mass are not limited to the internals of the furnace, but can also cause deformation of the supporting furnace structure, leading to a non-uniform furnace floor. Conventional tunnel designs utilize mortared joints to secure the bricks to one another, effectively turning the large number of small bricks into a small number of large wall sections. These wall sections act as a single body, and cannot accommodate any major dimensional change in the furnace floor. Deformation of the supporting furnace structure will therefore result in the failure of a conventional tunnel.
Differential thermal expansion occurs not only in situations with different design materials, but also across large sections of materials that are expected to act as a single body. Conventional tunnel design also uses fiber expansion joints roughly every 6-10 feet of wall length, with all of the building components in between adhered to one another with a refractory mortar. This refractory mortar also causes the wall sections to behave as a single body. No furnace has a completely uniform temperature distribution, however, and at some point, differential thermal expansion will occur across a wall section. The stresses imparted on the wall section are the same as those that cause thermal shock within a singularly body.
FEA has been performed to determine stress levels associated with a differential temperature from the top of a fully mortared 10 ft wall section to the bottom, where the fully mortared wall section was treated as a single body for the purposes of the analysis. The bottom of the wall section was 1925° F. and the top of the wall section was 1900° F., with a uniform temperature distribution in between. The FEA also included a simulated weight of the tunnel lids and gravity, but no other external forces. It was shown that the stress of the system exceeds the 500 psi HMOR of a standard refractory mortar. Since the mortar joints are the weakest point on the wall, they crack to alleviate the stress. The more cracking that occurs in the mortared wall, the smaller the wall sections become, and the lower the stresses become in any one section.
Properly accommodating for thermal expansion is one of the most difficult aspects of any thermal application design. Conventional tunnel designs use a different materials and designs for the tunnel lid and the tunnel base. Many tunnels have low density refractory or fiber insulation in the “base” area in between the wall supporting IFB columns. The tunnel lid can expand as much as ⅜ in, thereby pushing the tunnel walls apart, whereas the fiber insulation will not impart any expansion forces on the tunnel walls. The resulting trapezoidal shape is susceptible to buckling and collapsing. In certain situations, tunnels have been found at the conclusion of a furnace campaign to have alternative movement in the lateral direction. This is more commonly known as “snaking,” and is the result of the overall tunnel attempting to expand greater than the built-in allowance. This movement will crack the mortar, separate the walls from lids, and push the walls off of the IFB base; all of which lead to failure. While traditional tongue and groove brick design with a circular cross section (see, e.g., FIGS. 4 and 5) is somewhat effective in preventing lateral movement, this arrangement does will not sufficiently arrest buckling, as the rotation of one block relative to the block below it will separate the tongue from the groove, allowing a full system collapse (see, e.g., FIG. 15).
In addition to the above problems with the traditional wall design and components themselves, installing a conventional tunnel system requires a number of skilled labor positions that are becoming increasingly challenging to fill, particularly for temporary needs. This often creates a situation where the proper level of skilled labor is not available, and the overall quality of resultant installed tunnel system is compromised or the installation costs become higher than expected. In some instances, a conventional tunnel system has simply operated for the full amount of its originally projected life span, but due to short time frame of a turnaround schedule the tunnel cannot be fully repaired or replaced and must continue to perform for an extended campaign. The length of time and the high skill level required to install a conventional tunnel system therefore becomes a cause for the reliability issues. The full extent of damage that may be imparted to a tunnel system is often unknown prior to a turnaround, so a maintenance engineering crew has only a few weeks to examine, design, and implement repairs that are meant only to keep the tunnel system operational until the next turnaround, where this kind of repair can be attempted again. This is can be a very dangerous gamble for a plant, based on the long lead time and installation time associated with replacing the tunnels when a failure results in an unplanned outage.
The extended time frame and high level of skill required for installation and repairs imparts undesirable variability in quality output for conventional tunnel systems. Repairs that end up taking longer than the available window of plant turnaround time are not a viable option, and often result in an undesirably extended tunnel campaign. There is a strong desire to reduce the overall installation time and need for highly skilled labor in order to decrease this variability in quality. In some cases, conventional tunnel systems require overhead cranes to be installed to assist in the handling of the heavy tunnel lids.
In addition, in some cases, the end user may seek to add supplemental members to reinforce the tunnel wall structure and/or control the gas flow dynamics within the tunnel chamber in order to achieve improved efficiency, higher throughput or other specific results.
In order to control the gas flow, in the past, the tunnel design has been modified to exclude a brick or a half-brick from the array to provide the openings in the specific locations needed to achieve the objective, as discussed above. This conventional system, however, has much room for improvement. U.S. Pat. No. 8,439,102 discloses vector tiles that are used in conjunction with diffusor walls in reaction furnaces to control the gas flow direction, however, these vector tiles are not used in the sidewalls of hydrogen gas reformer flue gas tunnels, and since these vector tiles are cemented in place, they are not easy to install or fix in situ as the situation may demand.
To date, the prior art does not include any universally applicable refractory insert members that can be easily installed in the openings in the blocks in any location(s) desired by the end user to reinforce/strengthen the structural integrity of the tunnel and/or control the flow dynamics in any manner that is required for any particular type of application.